Additive Manufacturing Fatigue and Fracture IV: Toward Confident Use in Critical Applications: Property Prediction II
Sponsored by: TMS Structural Materials Division, TMS: Additive Manufacturing Committee, TMS: Mechanical Behavior of Materials Committee
Program Organizers: Nik Hrabe, National Institute of Standards and Technology; Steve Daniewicz, University of Alabama; Nima Shamsaei, Auburn University; John Lewandowski, Case Western Reserve University; Mohsen Seifi, ASTM International/Case Western Reserve University
Thursday 2:00 PM
February 27, 2020
Room: 10
Location: San Diego Convention Ctr
Session Chair: Nik Hrabe, National Institute of Standards and Technology; Jake Benzing, National Institute of Standards and Technology
2:00 PM Invited
Fatigue Behavior of Additive Manufactured Ni and Ti Alloys Through Coupled Modeling and In-situ Experiments: Michael Sangid1; 1Purdue University
The benefits of additive manufacturing have been well documented, but prior tothese materials being used in critical applications, the mechanisms for fatigue failure must be identified and the life of these materials must be determined for use in a design context. In this work, the fatigue behavior of selective laser melting IN718, 718Plus, and Ti-6Al-4V are investigated through detailed characterization and modeling efforts. Specifically, in situ loading is used to identify the strain evolution in these materials through high-energy x-ray diffraction and digital image correlation. Simulation-based predictions of material performance, including fatigue crack initiation have been developed as a means of accelerating the insertion of new materials by reducing the associated cost and time for materials development. The fatigue modeling framework is combined with uncertainty quantification, validation, and verification efforts of the model’s readiness level, in order to build trust in the predictive capabilities of the model.
2:30 PM
Maximizing the Fatigue Lifetime by Choosing the Best Build Orientation: Amin S. Azar1; Magnus Reiersen2; Even W. Hovig3; Mikkel M. Pedersen4; 1SINTEF; 2University of Oslo; 3Norwegian University of Science and Technology (NTNU); 4Aarhus University
Affected by the slicing conditions, the as-built surfaces of the additively manufactured components will have staircase effect. This indigenous surface topography creates local stress concentrations, directly influencing the fatigue lifetime. For this purpose, we have developed a new algorithm that shrink-wraps a finely triangulated surface on the designed part and analyzes the relative angle of all the surface patches to find the optimum part positioning provisions on the build platform towards minimization of surface stress concentrations. The model was calibrated by the experimental data through a novel white light interferometry (WLI) method. The entire procedure was implemented in the open source "Fatlab" toolbox to provide the first-in-class fatigue analysis code that takes the surface conditions into account when calculating the fatigue life of AM components. The calculated results show that the fatigue life of any as-built AM component can be improved by as much as 30% using this procedure.
2:50 PM
Micromechanical Modeling Driven Design of Fatigue Resistant Metal Additive Manufacturing Solutions: Anssi Laukkanen1; Matti Lindroos1; Tatu Pinomaa1; Tom Andersson1; Tomi Suhonen1; 1VTT Technical Research Center of Finland
Arguably the most critical performance dominating metal additive manufacturing (AM) materials and parts is their resistance to fatigue. We present an integrated computational materials engineering (ICME) methodology where the driver for establishing and designing the fatigue performance of AM solutions is micromechanical evaluation of fatigue. The approach is based on explicitly assessing fatigue critical material features at the microstructural scale, such as inclusions, porosity, surface roughness and followingly evaluating plastic slip to fatigue damage interactions. To support ICME workflows we include defect formation and rapid solidification simulations at powder bed scale, which provides the possibility to address fatigue resistance over the process-structure-properties-performance chain. The computational means are demonstrated with use cases consisting of maraging steels, high entropy and nickel based alloys. The results provide insights how ICME driven approaches can be used to design better fatigue resistant AM solutions and reliably control their respective properties in critical high end applications.
3:10 PM
Microstructure-based Fatigue Performance Analysis and Prediction of Additively Manufactured 316L Stainless Steel Subjected to Different Heat Treatments: Chola Elangeswaran1; Antonio Cutolo1; Charlotte de Formanoir1; Gokula Krishna Muralidharan2; Brecht Van Hooreweder1; 1KU Leuven; 23D Systems Leuven
A microstructural model for predicting fatigue behaviour is adapted for 316L stainless steel manufactured by laser powder bed fusion. An analytical model for dislocation accumulation and damage evolution during cyclic loading is employed. Experimental validation is performed with vertically built miniaturised fatigue samples in four microstructural conditions: as-built, stress relieved, fully annealed and hot isostatic pressed. The samples are fatigue tested at fully reversed axial tension-compression loads. Curve fitting constants are extracted from known conditions and extrapolated for predicted material conditions. The results show that the prediction reasonably agrees with the experimental values and offers scope of expanding the model to account for other surface and orientation-based influencing factors.
3:30 PM Break
3:50 PM Invited
Surface Morphology, Stress Concentrations, Micromechanical Modeling and Fatigue Life in 3D Printed Metals: Anthony Rollett1; Christopher Kantzos1; 1Carnegie Mellon University
The effects of surface morphology on stress concentrations and thus fatigue are discussed with respect to 3D printed metals. Both experimentally measured samples and synthetic structures generated via a fractal approach are analyzed to determine correlations with stress hot spots. Convolutional neural nets (CNNs) are applied to make the connection as robust as possible and also learn which features are critical through the use of viewports. CNNs are also useful for sorting fatigue fracture surfaces and creating a computed fractography is ongoing. Simulation of elasto-viscoplastic mechanical response using a spectral code is efficient for working with 3D image data and provides quantitative data for the studies. 3D printing offers a wide range of parameter space where, e.g., varying the contouring results in measurable variations in fatigue life. Deciphering the relative importance of variations in porosity content versus surface roughness variations is ongoing.
4:20 PM
Prediction of Fatigue Life of Flight-critical Metallic Components Fabricated by Additive Manufacturing: Xuesong Fan1; Baldur Steingrimsson2; Duckbong Kim3; Peter Liaw1; 1University of Tennessee; 2Imagars LLC; Portland State University; 3Tennessee Tech University
This abstract describes a comprehensive toolset for predicting the fatigue life of flight-critical metallic components fabricated by additive manufacturing (AM). Existing toolsets cannot predict how AM affects material properties of additively manufactured parts. Hence, we propose a machine learning (ML) framework for predicting fatigue properties of additively manufactured metallic components. The framework is a generalization of Statistical Fatigue Life model by one of the authors, and employs sophisticated, physics-based metallurgical prediction models. ML can help avoid inaccuracies in fitting traditional models to real-world fatigue data. ML can also account for all the sources that can impact fatigue life of AM components. In this work, we identify defects, inhomogeneity and unwanted features (DIUF) at macro, micro, and nano-levels in AM process and comparing different DIUFs between AM and casting. We then show how stress life (S/N curves) for Ti-6Al-4V can be accurately predicted, based on data already available from the literature.
4:40 PM
Coupling Damage Models to Multiscale Modeling of the Selective Laser Melting Process for Metals: Patcharapit Promoppatum1; Sabeur Msolli1; Jerry Siu Sin1; Mark Jhon1; 1Institute of High Performance Computing
Although multiscale modeling has been successfully used to aid design of parts, relatively little effort has been made to predict damage occurring during the build process. Herein, we propose that by coupling a micromechanics-based damage model to the inherent strain approach for residual stress, it is possible to predict the evolution of damage in the part during the build process. The numerical prediction was performed with a bottom-up approach, where the amount of porosity was firstly estimated. Subsequently, failure behaviors dictated by internal defects were determined from a micromechanical-based calculation. Coupling a damage behavior to a part scale calculation by imposing a cohesive layer at part-substrate interface, the failure at the interface can be assessed from the development of plastic strain and stress triaxiality during the build at part level. Ultimately, our model develops understanding of the relationships between scan strategy, porosity from unmelted powder, and the ductile fracture process.